Unverified50% confidenceOpinionTime unknown
智能筛选回归用例Agent能将回归测试的用例数量压缩到全量的20%-30%，同时保持95%以上的缺陷发现率
1
Sources
50%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
Unverified实验采用8:2比例划分数据，20%作为内部测试集，训练集内部使用5折交叉验证，并移除重复记录以避免数据泄漏71% similarUnverified在标准行人检测数据集（如MOTChallenge）上训练的模型，在俯视视角下的检测精度（mAP）通常下降40%-60%68% similarUnverifiedIn the experimental design, the AI used an 8:2 split to reserve 20% as an independent test set, and applied 5-fold cross-validation on the 80% training set to avoid data leakage.68% similarUnverified运行10次测试置信区间宽度约±31%，100次收窄至±10%，400次进一步收窄至±5%，推荐关键测试用例至少运行20-50次68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/26595API
curl https://kongchang.com/api/v1/knowledge/claims/26595MCP
get_claim(id=26595)